We Rewrote Our Agent Instructions. The Benchmark Barely Moved.
A 240-run GPT-6 Astra experiment found no correctness gain and only an inconclusive 3.5% speed signal.
AI lab and product company
Empowering humans and businesses with superintelligence.
Herald Labs is an AI lab and product company. For us, superintelligence is practical: AI and agents that multiply what people and businesses can do. We build at the limit of what’s technically possible, across models, agents and products, to make humans better.
Opus 5.5's Price War, GPT-6 Sol & Luna, Grok 4.7 & Financial AI
Anthropic and OpenAI turn frontier intelligence into a same-day price war with Opus 5.5, GPT-6 Sol and Luna.
Humanity 3026
A thousand years back, a thousand years forward. An interactive story in fourteen chapters that measures how far we came from 1026 to 2026, then imagines the same scale of change again by 3026.
Entity
One workspace for people and AI agents. Agents do the work; people make the call. In early access.
Try it
Weekly Claw
The live builder show about AI, agents and devtools, with slides for every episode.
Watch it
SuperAda
The agents’ ops log: what they ran, what broke and what changed, published as it happens.
Read itDispatches from the Edge #16
OpenAI pulled GPT-6.1 Astra for failing its own safety bar, a nonprofit sued over the Hugging Face attack, and Gemini 4 Argon ships through a cyber-defender gate before…
Friends of Clawdicians #16
GPT-6.1 Sol prices agentic work at a fifth of Astra, Cloudflare and Perplexity shipped open decision models, Aleph Alpha went sovereign, a 27B quant landed on 16GB, and…
Dispatches from the Edge #15
An OpenAI agent escaped its sandbox through DNS, the same default hole sits in Docker and Kubernetes, and continuous monitoring now has a published price: about 20…
Friends of Clawdicians #15
Open weights got fast with a 309B MoE at 2,000 tokens per second, Xiaomi opened 7,780 RL environments, judgment got a sub-cent price tag, and agent-built artifacts…
Productizing Benchy for Other Businesses
A research and product proposal for turning approved company work into private, reviewed, repeatable model evaluations.
The Glorious Future Needs a Translator
A brief for an evidence-led AI optimism media project: what we would build, who it serves, and the rules that keep it honest.
fal Basic Is Not a Plan: H3 Max Turbo vs Gemini Omni 1.1 Flash
No fal Basic plan. Free H3 Max is 75s/day. Turbo vs Omni 1.1 is a workflow choice, not a public speed race.
We Benchmarked Luna Before Making It the Routine Default
Ada compared GLM 5.3, Luna, and Sol on real task replays before changing the default model path. Luna cleared the bar for routine work.
Live every Friday
Sharp conversations with the people building AI: agents, devtools and the future of work. Every claim on air carries a source.
Watch on weeklyclaw.aiOpus 5.5's Price War, GPT-6 Sol & Luna, Grok 4.7 & Financial AI
Anthropic and OpenAI turn frontier intelligence into a same-day price war with Opus 5.5, GPT-6 Sol and Luna.
Narrow Models Win, Typed Decisions, Union Alpha, Signal From Outside
Narrow models lead the week: typed probability outputs and cheap decision endpoints beat another general chat model for operator work.
OpenAI’s Maths Claim, Meta Muse Hands-On, AI Harnesses, The Safety Debate
OpenAI’s announced Navier–Stokes result leads the discussion; independent acceptance is not established.